The overall aim of this project is to determine, using 31-Phosphorous (31P) and proton (1H) nuclear magnetic resonance (NMR) spectroscopy, if brain metabolic markers exist that will distinguish Alzheimer's disease from dementia associated with multiple sites of subcortical ischemia. A five year longitudinal study will investigate clinical, imaging (magnetic resonance imaging, MRI), metabolic (in vivo 31P and 1H NMR) and neuropathologic data in selected patients with multiple subcortical ischemic dementia (MSID), Alzheimer's disease (AD), or mixed AD and vascular disease. and in nondemented elderly controls. The study aims to characterize changes in brain energy phosphate, phospholipid metabolism, pH, free magnesium, N-acetyl aspartate, and cholinated compounds that might accurately predict clinical and pathologic diagnoses of AD, MSID, and mixed MSID/AD. Further, two hypotheses will be tested to explain the shift towards the high energy phosphate end of the phosphorus spectrum previously reported in MSID. One hypothesis attributes the shift to a decreased utilization of molecular energy in superficial brain regions disconnected from neuronal input by subcortical lesions. The other hypothesis attributes the shift to increased astrocytic reactivity in cortical regions that have been subjected to mild levels of ischemia. The first hypothesis predicts that the extent of the high energy phosphate shift should be correlated with the extent and location of the subcortical white matter hyperintensities observed on MRI. The second hypothesis predicts that MSID patients with a shift towards the high energy phosphate end of the phosphorus spectrum will have microcavities and increased glial immunoreactivity in cortical regions where spectra are obtained, even though these areas appeared to be normal on MRI.